Expect an exponential

Calculus Level 5

Let P ( n ) = a n P(n)=a^n . Define f n ( x ) f_n (x) as a least-degree polynomial that passes through the coordinates ( x , y ) = ( 1 , P ( 1 ) ) , ( 2 , P ( 2 ) ) , ( 3 , P ( 3 ) ) , , ( n , P ( n ) ) (x,y) = (1, P(1)), (2,P(2)) , (3,P(3)) , \ldots , (n, P(n)) Define the expected function, P e ( n ) Pe(n) to be equal to f n 1 ( n ) . f_{n-1} (n). lim n P ( n ) P e ( n ) P ( n ) = b \lim_{n\to\infty} \dfrac{P(n) - Pe(n)}{P(n)}=b How many possible finite values exist for b b ?

inspiration

0 1 4 infinitely many 3 2

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Mark Hennings
Jun 21, 2018

If a = 0 a=0 then P ( n ) = 0 P(n) = 0 for all n 1 n \ge 1 , so that f n ( X ) 0 f_n(X) \equiv 0 for all n 1 n \ge 1 and hence P e ( n ) = 0 Pe(n) = 0 for all n 1 n \ge 1 . The ratio P ( n ) P e ( n ) P ( n ) \frac{P(n) - Pe(n)}{P(n)} is not defined. If a = 1 a=1 then P ( n ) = 1 P(n) = 1 for all n 1 n \ge 1 , so that f n ( X ) 1 f_n(X) \equiv 1 for all n 1 n \ge 1 , so that P e ( n ) = 1 Pe(n) = 1 for all n 1 n \ge 1 , and hence P ( n ) P e ( n ) P ( n ) = 0 \frac{P(n) - Pe(n)}{P(n)} = 0 for all n 1 n \ge 1 . Thus b = 0 b=0 is one possible value.

Let us now assume that a 0 , 1 a \neq 0,1 . Lagrangian interpolation gives us that f n ( X ) = u = 1 n a u 1 v n v u X v u v = u = 1 n a u ( 1 ) n u ( u 1 ) ! ( n u ) ! 1 v n v u ( X v ) = 1 ( n 1 ) ! u = 1 n ( 1 ) n u ( n 1 u 1 ) a u 1 v n v u ( X v ) \begin{aligned} f_n(X) & = \; \sum_{u=1}^n a^u \prod_{{1 \le v \le n} \atop {v \neq u}} \frac{X-v}{u-v} \; = \; \sum_{u=1}^n \frac{a^u}{(-1)^{n-u}(u-1)!(n-u)!}\prod_{{1 \le v \le n} \atop {v \neq u}} (X-v) \\ & = \; \frac{1}{(n-1)!} \sum_{u=1}^n (-1)^{n-u}\binom{n-1}{u-1}a^u \prod_{{1 \le v \le n} \atop {v \neq u}}(X-v) \end{aligned} and so f n ( X ) f_n(X) has degree n 1 n-1 and leading coefficient α n = 1 ( n 1 ) ! u = 1 n ( 1 ) n u ( n 1 u 1 ) a u = 1 ( n 1 ) ! u = 0 n 1 ( 1 ) n 1 u ( n 1 u ) a u + 1 = 1 ( n 1 ) ! a ( a 1 ) n \begin{aligned} \alpha_n & = \; \frac{1}{(n-1)!}\sum_{u=1}^n (-1)^{n-u}\binom{n-1}{u-1}a^u \; = \; \frac{1}{(n-1)!} \sum_{u=0}^{n-1} (-1)^{n-1-u}\binom{n-1}{u}a^{u+1} \\ & = \; \frac{1}{(n-1)!}a(a-1)^n \end{aligned} On the other hand we know that f n + 1 ( X ) f n ( X ) f_{n+1}(X) - f_n(X) is a degree n n polynomial that vanishes at 1 , 2 , . . . , n 1,2,...,n , and so it is clear that f n + 1 ( X ) f n ( X ) = α n + 1 ( X 1 ) ( X 2 ) ( X n ) f_{n+1}(X) - f_n(X) \; = \; \alpha_{n+1}(X-1)(X-2)\cdots (X-n) Putting X = n + 1 X = n+1 we see that f n + 1 ( n + 1 ) f n ( n + 1 ) = α n + 1 n ! P ( n + 1 ) P e ( n + 1 ) = a ( a 1 ) n \begin{aligned} f_{n+1}(n+1) - f_n(n+1) & = \; \alpha_{n+1} n! \\ P(n+1) - Pe(n+1) & = \; a(a-1)^n \end{aligned} so that X n = P ( n + 1 ) P e ( n + 1 ) P ( n + 1 ) = ( a 1 a ) n X_n \; = \; \frac{P(n+1) - Pe(n+1)}{P(n+1)} \; = \; \left(\frac{a-1}{a}\right)^n If a > 1 2 a > \tfrac12 then 1 < a 1 a < 1 -1 < \tfrac{a-1}{a} < 1 and hence b = 0 b=0 . If a = 1 2 a=\tfrac12 then X n = ( 1 ) n X_n = (-1)^n and so b = lim n X n b = \lim_{n \to\infty}X_n does not exist. If 0 < a < 1 2 0 < a < \tfrac12 then a 1 a < 1 \tfrac{a-1}{a} < -1 , and so the sequence X n X_n fails to converge. If a < 0 a < 0 then a 1 a > 1 \tfrac{a-1}{a} > 1 and so the series X n X_n fails to converge. Thus there is only 1 \boxed{1} possible value of b b , namely b = 0 b=0 , which is achieved when a > 1 2 a > \tfrac12 .

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...